Re: Scalable cluster - Mailing list pgsql-general

From Mason S
Subject Re: Scalable cluster
Date
Msg-id CA+rR5x0XueiEe3e+j7sDgM=dPtKQprfuj37q6NyeT4g3pjPK8Q@mail.gmail.com
Whole thread Raw
In response to Re: Scalable cluster  (Daniel de Oliveira Mantovani <daniel.oliveira.mantovani@gmail.com>)
List pgsql-general


On Mon, Mar 4, 2013 at 1:01 PM, Daniel de Oliveira Mantovani <daniel.oliveira.mantovani@gmail.com> wrote:
just fixing:

like Netezza which has full implementation to do it.*

On 4 March 2013 14:59, Daniel de Oliveira Mantovani
<daniel.oliveira.mantovani@gmail.com> wrote:
> One question about Postgres-XC, can it distribute one single query in
> all nodes ?
>
> Example,
> Nodes: Node1,Node2,Node3
> Query:"SELECT sum(foo) FROM bar group by foo"

XC handles queries like this in parallel to a certain extent.

A plain old SELECT SUM(foo) FROM bar will get the sum on all of the nodes, then sum it up again at the coordinator.

Depending on your table distribution scheme, XC will do a pretty good job at pushing down joins so that they occur locally.  For example, one table is distributed, and a second is replicated on all nodes. Any join between the two tables can occur locally on each data node in parallel.

You will experience slow query times if you have to join data from one node with data from another node. It all gets shipped to the coordinator for joining. 

If you are after query parallelism for a data warehouse and want to be PostgreSQL-based, then I recommend a project called Stado. If you are looking for mainly write scalability where the workload is largely singleton type of statements, then I recommend Postgres-XC.

Yet another solution you can consider is PL/Proxy, which was developed by Skype.



>
> (Obvious: Your data has to be shared between the nodes)
>
> Query -> Compiler (in the pool)-> "share the query between the nodes"
> -> merge -> return
>
> I know tools to do it, but don't have full implementation like Netezza.
>
> Thank you.
>
> On 4 March 2013 13:36, Koichi Suzuki <koichi.szk@gmail.com> wrote:
>> Version 1.1 will be out in this June time frame, with online node
>> addition/removal, Trigger and improved planner.  I'm more than happy
>> if you evaluate XC.
>>
>> Regards;
>> ----------
>> Koichi Suzuki
>>
>>
>> 2013/3/4 Greg Jaskiewicz <gryzman@gmail.com>:
>>>
>>> On 4 Mar 2013, at 13:21, Vincent Veyron <vv.lists@wanadoo.fr> wrote:
>>>
>>>>
>>>> There is this :
>>>>
>>>> http://wiki.postgresql.org/wiki/Postgres-XC
>>>>
>>>>
>>>> [Disclaimer : I can't tell whether it's usable or not; I just know they
>>>> exist]
>>>>
>>>
>>>
>>> Well, I know of its existence too. Question is how production ready is it. And also how far off the 9.2 line is it.
>>>
>>>
>>>
>>> --
>>> Sent via pgsql-general mailing list (pgsql-general@postgresql.org)
>>> To make changes to your subscription:
>>> http://www.postgresql.org/mailpref/pgsql-general
>>
>>
>> --
>> Sent via pgsql-general mailing list (pgsql-general@postgresql.org)
>> To make changes to your subscription:
>> http://www.postgresql.org/mailpref/pgsql-general
>
>
>
> --
>
> -dom
>
> --
>
> IBM - Business Analytics Optimization Consultant
> Daniel Mantovani +5511 8538-9897
> XOXO



--

-dom

--

IBM - Business Analytics Optimization Consultant
Daniel Mantovani +5511 8538-9897
XOXO


--
Sent via pgsql-general mailing list (pgsql-general@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-general

--
Mason Sharp

StormDB
The Database Cloud
Postgres-XC Support and Services 

pgsql-general by date:

Previous
From: Maciek Sakrejda
Date:
Subject: Re: [HACKERS] Floating point error
Next
From: Daniel Farina
Date:
Subject: Re: [HACKERS] Floating point error